Psychometric Analyses of the Leyton Obsessional Inventory in Patients With Obsessive-Compulsive and Other Anxiety Disorders


This study evaluated the psychometric properties of the Leyton Obsessional Inventory (LOI). In particular, internal consistency and intercorrelation of subscales were examined, as well as convergent, divergent, and discriminative validity. In a sample of 77 patients with primary DSM-III-R diagnoses of obsessive-compulsive disorder (OCD; n = 18) or alternative anxiety disorders (n = 59), internal consistency of the 4 LOI subscales (Symptom, Trait, Resistance, Interference) was strong, and subscale intercorrelations suggested the presence of a single construct. Analyses of convergent and divergent validity, including the SCL-90 and EPI subscales, supported the specificity of the LOI. In addition, OCD and non-OCD patients differed significantly in Symptom, Interference, and Resistance scores. Discriminant function analyses revealed that Interference and Symptom subscales were the strongest predictors of group membership, together accounting for 57% of the variance. Classification data indicated that 88% of patients in the OCD and non-OCD groups were classified correctly according to LOI scores. A LOI Interference cutoff score of 15 may be optimal for identifying OCD and non-OCD groups. Implications of the data for use of the LOI in assessing OCD are discussed.

Publication Title

Psychological Assessment